The Four Things That You Lose with Scale Computing HC3
-
@dafyre said:
@craig-theriac : Has Scale ever considered selling single nodes for small businesses?
Two is one, and one is none.
It would need to be at least 2 nodes.
-
@dafyre i cannot confirm nor deny
@aaronstuder let me know / send me a DM if you want to jump on a call to go through our roadmap. I'd love to learn more about your requirements for a lower end solution. -
@aaronstuder said in The Four Things That You Lose with Scale Computing HC3:
@dafyre said:
@craig-theriac : Has Scale ever considered selling single nodes for small businesses?
Two is one, and one is none.
It would need to be at least 2 nodes.
Ideally, you are correct. But if I only need a single server, why plunk down $25k for two Scale nodes when I can buy a Dell of decent quality and similar specs and install $hypervisor myself? ... Granted an SMB isn't likely to think like that, lol.
Don't get me wrong -- I've enjoyed working with the Scale team and their systems! That was just a burning question always in the back of my mind.
-
@dafyre said:
Ideally, you are correct. But if I only need a single server, why plunk down $25k for two Scale nodes when I can buy a Dell of decent quality and similar specs and install $hypervisor myself? ... Granted an SMB isn't likely to think like that, lol.
Don't get me wrong -- I've enjoyed working with the Scale team and their systems! That was just a burning question always in the back of my mind.
It's in the back of my mind too. Hoping @craig.theriac could tell us why there better
-
@dafyre said in The Four Things That You Lose with Scale Computing HC3:
@aaronstuder said in The Four Things That You Lose with Scale Computing HC3:
@dafyre said:
@craig-theriac : Has Scale ever considered selling single nodes for small businesses?
Two is one, and one is none.
It would need to be at least 2 nodes.
Ideally, you are correct. But if I only need a single server, why plunk down $25k for two Scale nodes when I can buy a Dell of decent quality and similar specs and install $hypervisor myself? ... Granted an SMB isn't likely to think like that, lol.
Don't get me wrong -- I've enjoyed working with the Scale team and their systems! That was just a burning question always in the back of my mind.
That's our current SMB setup - several 1/2U dell servers w/ hyperv. Not sure where / how Scale would help us but honestly that's probably because I've yet to read their specs etc.
-
@MattSpeller Just one server? We do the same, but 2 servers at each site.
-
@aaronstuder said in The Four Things That You Lose with Scale Computing HC3:
@MattSpeller Just one server? We do the same, but 2 servers at each site.
We're at 5 at primary, 2 at secondary, 1 at a remote
Could probably replace the 5 with two current nicely spec'd ones but $$$$$$$$$$$$
-
@MattSpeller You sound like us
-
@aaronstuder said in The Four Things That You Lose with Scale Computing HC3:
@MattSpeller You sound like us
Broke SMB sounds similar the world over heheh
#not4profitlife
-
1 & 2. Sure.
3 - Well Hyper-v and Xen-server are free.
4 - Is an all your eggs in one basket model. Which can be risky too.
-
@Breffni-Potter said in The Four Things That You Lose with Scale Computing HC3:
1 & 2. Sure.
3 - Well Hyper-v and Xen-server are free.
4 - Is an all your eggs in one basket model. Which can be risky too.
4... Backup, backup, backup!
-
@dafyre said
4... Backup, backup, backup!
Backups funnily enough don't work for.
- Hardware availability. Can I get a replacement node? If so how quickly.
- If the software part of the hyper-visor all in one magic box breaks and I lose all 3 nodes, what then.
- Scale takes away a lot of the tech from the technician to make it easier which is good but when that tech fails, what can you do.
Now, I like the Scale model but it's still a magic box in the corner.
-
@Breffni-Potter said in The Four Things That You Lose with Scale Computing HC3:
@dafyre said
4... Backup, backup, backup!
Backups funnily enough don't work for.
- Hardware availability. Can I get a replacement node? If so how quickly.
- If the software part of the hyper-visor all in one magic box breaks and I lose all 3 nodes, what then.
- Scale takes away a lot of the tech from the technician to make it easier which is good but when that tech fails, what can you do.
Now, I like the Scale model but it's still a magic box in the corner.
Yeah, that is a perfectly valid point... and I think that is why their starter kits are more than one node, lol.
-
@dafyre said
Yeah, that is a perfectly valid point... and I think that is why their starter kits are more than one node, lol.
Yes, but even if I buy 500 nodes, the software is the single point of failure.
Look at Microsoft Azure.
They've not had hardware problems, Not lost data centres, their issues have been software based.
-
@Breffni-Potter said in The Four Things That You Lose with Scale Computing HC3:
@dafyre said
Yeah, that is a perfectly valid point... and I think that is why their starter kits are more than one node, lol.
Yes, but even if I buy 500 nodes, the software is the single point of failure.
Look at Microsoft Azure.
They've not had hardware problems, Not lost data centres, their issues have been software based.
But you could still run into that with VMware, or Xen, or Hyper-V, or KVM, or any piece of software. I'm glad to say, I've never seen an issue with the Scale systems take out more than one node (and that was a hardware failure).
-
@Breffni-Potter said in The Four Things That You Lose with Scale Computing HC3:
@dafyre said
4... Backup, backup, backup!
Backups funnily enough don't work for.
- Hardware availability. Can I get a replacement node? If so how quickly.
- If the software part of the hyper-visor all in one magic box breaks and I lose all 3 nodes, what then.
- Scale takes away a lot of the tech from the technician to make it easier which is good but when that tech fails, what can you do.
Now, I like the Scale model but it's still a magic box in the corner.
It's not, really. Because it runs on KVM you can always take a backup and restore to any hardware with KVM without doing a P2V and you can always to a P2V to disparate hardware. So there isn't any lock in.
The "software point of failure" is common to all clustered solutions. The same applies to XenServer, Hyper-V and ESXi - if you cluster them, they become a single computer. It's the Catch-22 of HA solutions.... by removing all of the single points of failure, you create a new one. But the hope is that the new one is dramatically less likely to fail than the ones that it replaced; and generally that is true.
-
@dafyre said
But you could still run into that with VMware, or Xen, or Hyper-V, or KVM, or any piece of software.
Yes but there is a world of difference between those 3 and Scale in sheer volume of installs alone.
How many Scale competent guys can I get versus someone who works in any of the above?
-
@dafyre said in The Four Things That You Lose with Scale Computing HC3:
@Breffni-Potter said in The Four Things That You Lose with Scale Computing HC3:
@dafyre said
Yeah, that is a perfectly valid point... and I think that is why their starter kits are more than one node, lol.
Yes, but even if I buy 500 nodes, the software is the single point of failure.
Look at Microsoft Azure.
They've not had hardware problems, Not lost data centres, their issues have been software based.
But you could still run into that with VMware, or Xen, or Hyper-V, or KVM, or any piece of software. I'm glad to say, I've never seen an issue with the Scale systems take out more than one node (and that was a hardware failure).
They do an extreme amount of testing. Unlike most solutions that only test compatibility, they test the full stack of hardware configurations, software versions, and even firmware in every permutation so you know that you matter what you are running, it's a tested configuration.
-
@Breffni-Potter said in The Four Things That You Lose with Scale Computing HC3:
@dafyre said
But you could still run into that with VMware, or Xen, or Hyper-V, or KVM, or any piece of software.
Yes but there is a world of difference between those 3 and Scale in sheer volume of installs alone.
Is that really a factor, though? Scale only has a couple configurations to test, each of those literally have millions of configurations that they don't test. So when thinking of the scale for testing, Scale actually has the advantage.
-
@scottalanmiller said
It's not, really. Because it runs on KVM you can always take a backup and restore to any hardware with KVM without doing a P2V and you can always to a P2V to disparate hardware. So there isn't any lock in.
Ok, but the Scale model says you don't need to know how to do that. It's entire selling point is simplicity and ease of use.
